h1

Lib Dem incumbency would be overwhelmed on current polling

May 31st, 2014

Their current national figures would see losses on the same scale as the local elections

Cockroach-like.  That was Tim Farron’s description of Lib Dems’ resilience in withstanding a hostile climate.  The inference was that no matter how tough things might be across the country, where they have elected representatives, their vote would hold firm enough.

He had a point: Lib Dem MPs and councillors have in the past proven notoriously difficult to shift due to local campaigning, popularity and hard work and their being transfer-friendly to tactical votes.  Is it still true though?  The recent constituency polling Lord Oakeshott commissioned suggested strongly not, though it should be noted that the methodology ICM used was unfavourable to his former party.

A better case study was the local elections: for every four seats held last week, three were lost.  Not all of them will have been in constituencies the Lib Dems represent at Westminster or even ones they’re competitive in but by definition they must have been locally strong in the relevant wards and yet more than 40% defended still went down.  The same rate of loss at a general election would leave the Yellows with 33 MPs.

In some ways, that wouldn’t be at all bad: to lose half your vote but keep more than half your MPs would be good going, especially under FPTP.  After all, 33 MPs is considerably more than the Liberal-SDP Alliance won in the 1980s on close to three times the current polling.

Incumbency can only hold out against the tide so far though: there comes a point when there simply aren’t enough votes to go round to return large numbers of MPs: UNS cannot apply at the bottom end because you can’t lose (say) 14% in a seat where you only had 8% to start with.  Of itself, that means that more have to come from the middle or top to make up the difference.  If we further assume that the top end also outperforms because of incumbency, that implies a tremendous and probably unrealistic collapse in the middle.

To give an example, suppose the Lib Dems poll 9% at the election.  On the same turnout as 2010, that would reduce their votes received from 6.8m to 2.7m (not that the turnout really matters but keeping it the same removes one variable).  If they polled 30% in twenty seats, 35% in another twenty and 40% in a further twenty, that alone would be close to a million votes – and those shares might well see them win the 33 mentioned above.  It would, however, leave just 1.7m to be spread across the 570 other constituencies at an average of just 6.4%.  With presumably other seats where the Lib Dems have a reasonable showing, it’d imply near-extinction levels across many and lost deposits in about half.

If Polling into single figures is inconsistent with keeping the substantial majority of seats presently held but the question is which side which part will prove to be false.  One clue could be provided if the polling companies routinely asked people to think about how they’d vote in their constituency.  We know this can have an effect on responses and as everyone does vote in their constituency, there’s no real reason not to phrase it like that.  If there is an incumbency bonus, doing so might bump up the polls and square the circle.  On the other hand, if the polling is right – and present figures lie smack bang between the Lib Dems’ actual scores in the locals and Euros, just as you’d expect – then the Yellows are in for big losses.

David Herdson